Wood floor surface defect detection method based on unsupervised deep learning

Abstract

The invention relates to wood floor surface defect detection method based on unsupervised deep learning, which comprises the steps of 1) collecting samples, wherein the samples include non-defective product samples and defect samples, and the numbers of the non-defective product samples and the defect samples are both greater than 5000; 2) carrying out image noise preprocessing on the collected samples by adopting a Gaussian filtering mode; 3) carrying out sample expansion on the samples; 4) establishing a network model, and generating a deep learning operation library; and 5) detecting a to-be-detected product, preprocessing the to-be-detected product in the same way as the step 2), inputting the preprocessed to-be-detected product into the deep learning operation library generated in thestep 4) for judgment, and giving a judgment result of a non-defective product or a defective product. The problem that existing wood floor surface layer grain changes and the like cannot be detectedin the prior art is solved, the purpose of replacing manual detection can be achieved, effective detection of surface defects in the wood floor industry is achieved, and the industrial production quality standard is met.

Description

technical field [0001] The invention relates to the technical field of visual inspection, in particular to an unsupervised deep learning-based detection method for wooden floor surface blemishes. Background technique [0002] In order to ensure that the products meet the quality requirements when the floor is delivered, the staff must inspect all the finished products to ensure that there are no surface defects in the products that affect the integrity and aesthetics of the products. A large number of defects are missed, and the efficiency of manual inspection is low. In response to this problem, machine vision inspection has been initially applied, and a platform equipped with an optical imaging and inspection system is used to detect and judge foreign object defects on products. [0003] The detection methods currently applied to machine vision are mainly: defect comparison detection and deep learning algorithm, and the existing problems are: [0004] 1. The existing tra...

AI Technical Summary

Problems solved by technology

[0008] The technical problem to be solved by the present invention is to provide a method for detecting wooden floor surface flaws based on unsupervised deep learning, which solves the problem of floor surface flaw detection and achieves the effect of replacing manual surface flaw detection with visual detection

Images